Altered brain network centrality in patients with mild cognitive impairment: an fMRI study using a voxel-wise degree centrality approach

Introduction

Alzheimer’s disease (AD), a progressive and irreversible neurological disorder, is characterized by memory loss, aphasia, agnosia, visual and spatial skill impairment, executive dysfunction, and personality and behavioral changes. For these reasons, it is a common cause of dementia. It has a long preclinical period with pathophysiological changes occurring 15 to 20 years before the onset of symptoms [1, 2]. The prevalence of AD increases exponentially with age, as reflected by its frequent occurrence in the elderly [2]. No effective treatment is available for AD and late interventions often result in treatment failure. Early detection and diagnosis of AD require efficient humoral and imaging markers, which are still in the developing stage. Medical imaging, including diffusion tensor imaging, fluorodeoxyglucose positron emission tomography, magnetic resonance imaging (MRI), and other methods can detect brain lesions in patients with AD. Monitoring of neurodegenerative changes in the central nervous system can provide direct information on disease progression [3]. The early stage of AD involves damage to the hippocampus and entorhinal cortex, whereas diffuse brain atrophy is associated with the late stage of AD [4].

Mild cognitive impairment (MCI) refers to a minor but noticeable decline in cognitive abilities. It is a heterogeneous clinical entity with multiple causes. The European Consortium on Alzheimer’s Disease Working Group on MCI describes a neuroimaging- and biomarker-based three-step process to identify MCI patients with incipient or prodromal AD [5, 6]: 1) decline in cognitive functions during the past year, 2) although patients’ clinical assessment shows cognitive impairment, they do not suffer from dementia and no major impact on daily life, and 3) identification of MCI subtypes responsible for AD.

In addition, amyloid (Aβ) accumulation in the brain of early-stage AD patients with MCI and no other clinical symptoms assist in the identification.

Although functional magnetic resonance imaging (fMRI) is an expensive and time-consuming brain-imaging method, it objectively assesses brain functions.

For instance, voxel-wise degree centrality (DC), an fMRI-based method, provides major insights into the functional connectivity of the whole brain and has been widely used to assess the pathophysiological mechanism of several diseases [7]. In contrast with other methods, such as amplitude of low-frequency fluctuations (ALFF) [8] and regional homogeneity (ReHo) [9], DC does not involve defining regions of interest (ROI) and evaluate the connection strength of the whole human brain at the voxel level [10]. We used DC to evaluate changes in brain functional connectivity in patients with MCI and any relationship between these changes and clinical symptoms.

Materials and Methods

Results

Discussion

DC is a reliable, resting-state fMRI method to measure brain activity and connectivity. It has been successfully applied to study several neurogenic and ophthalmological diseases (Table 3). While assessing the brain activity in patients with MCI, we found the DC values to be lower than those in controls in the left ITG but relatively high in the left SMA and right fusiform gyrus. Furthermore, this alteration occurred earlier than the changes in the eyes (Figure 5). In addition, the left ITG showed a positive correlation with MMSE in MCI patients and a negative correlation with disease duration (Figure 6).

The temporal lobe is located below the lateral fissure and processes vision, olfaction, sensation, and memory. The relatively low DC values in the left ITG in MCI patients suggest impaired MCI. We used the available data to study the association between MMSE and disease duration. The DC values in the left ITG decreased with duration (Figure 3A) but were positively correlated with MMSE scores (Figure 3B). These results show that MCI develops over time and affects the cognitive process.

In addition, the data revealed higher Sβ100 in the MCI group than in the HC group. Previous studies have reported this phenomenon in AD [12] but not in MCI. Thus, high S1β00 is related to the development of AD and could be a potential biomarker to identify MCI subtypes leading to AD.

Abnormal phosphorylation of tau protein hinders cell–cell connections, disrupts the axons of neurons, and causes pathological damage to the body. Accumulation of Aβ greatly accelerates the normal spread of tau with aging through neuronal communication pathways [13]. This phenomenon is observed in patients with AD and manifested as plaques formed by β-amyloid and tangles of neurofibrils within nerves, severely impairing brain function. In addition, it is speculated that plaque accumulation initiates during MCI development, causing low DC values of left ITG. The temporal lobe, from which these lesions emanate [14–17], coordinates social, emotional, language functions, and long-term episodic memory and is associated with autism and language deficits [18]. Dehaene et al. [19] proposed the concept of a “local combined detector” and suggested the function of ITG in information transmission and language learning [20].

In addition, ITG is involved in common object perception. A study [21] found reduced ITG and fusiform gyrus function in autistic patients during face recognition, a finding consistent with that of our study. However, the present results are inconsistent with those reported previously [22]. We attribute these inconsistencies to differences in age between patients in different studies.

Aβ accumulates in the retinas of patients with MCI, probably before accumulating in the brain [23]. In addition, cataract patients are 1.43 times more likely to have AD than normal individuals [24]. Although we did not study whether the vision was impaired in patients with MCI, AD can be complicated by ocular diseases, causing visual impairment [25]. Thus, Aβ in the eyes can serve as an early clinical indicator for AD prior to clinical symptoms.

Fusiform gyrus, a region in the occipitotemporal lobe, participates in advanced visual processing, especially target recognition and categorization of complex stimuli (such as human faces). People with autism have cognitive defects associated with the fusiform gyrus, including impaired facial processing and abnormalities in the fusiform gyrus. Patients with MCI are at an increased risk of depression and anxiety [26, 27]. Although the enhanced activity of the right fusiform gyrus has been found in patients with depression, a study showed it not be associated with these symptoms [28]. We suspect that increased DC values observed in the right fusiform gyrus in MCI patients compared to HC patients could be potentially associated with the occurrence of depressive symptoms.

SMA is located on the medial side of the primary motor cortex and the medial part of the premotor cortex, roughly equivalent to Brodmann’s area 6 (BA6). It is composed of two regions, pre-SMA (rostral side) and SMA proper (caudal part). Although SMA is more active than the cerebral cortex during simple movements such as limb movements, it is associated with complex motor memory as well. It coordinates the temporal organization of movements, especially the execution sequence of multiple movements [29]. In addition, SMA proper is implicated in the planning of complex movements of learning (believed to be internally driven than by visual cues) and the coordination of movements involving both hands. Pre-SMA participates in the learning of new motion sequences. A previous study reported increased interhemispheric functional connectivity with Aβ accumulation in SMA. In addition, we reported a higher DC value of SMA in the MCI group than in the HC group. These findings suggest a compensatory enhancement through the recruitment of other networks for non–amyloid-dominant dementia [12].

This study has certain limitations. First, the number of patients was limited, resulting in low reproducibility of results and overestimation of effects. Second, inevitable individual differences could have impacted the experimental results. Finally, patients’ daily lifestyles will inevitably affect the development of MCI.

Conclusions

MCI can lead to anomalies in specific regions in the patient’s brain and induce a variety of symptoms, such as affective expression disorders and retinopathy. These findings preliminary validate the relevant theories, such as brain abnormalities due to Aβ accumulation in AD, and provide a basis to further investigate neural changes in MCI and AD. We believe these findings will assist in the early diagnosis of MCI and predict the development of MCI into AD.